Website founded by Milan Velimirović in 2006
15:21 UTC
| |
MatPlus.Net Forum Internet and Computing Problem solvers suitable for conditional problems? |
|
|
|
You can only view this page!
| | (1) Posted by [Tuesday, Oct 15, 2019 10:19]; edited by [19-10-15] | Problem solvers suitable for conditional problems? Are there any problem solvers that can provide *all* solutions that meet the stated stipulation, and not give up as soon as a short solution is found?
I've looked at Popeye and dchest/Winchest, but they don't seem to be able to handle this -- though I may have missed some obscure option.
Pure GUI-based solvers are of less interest, unless they can save the solutions in a reasonably open and processable format.
This is not 'normal' use for problem solvers, but I think it could be useful for testing conditional problems.
Example:
42, Lichtenstein: Der Schachkünstler
(= 9+6 )
1r2N3/4NQ2/2p5/qp2kP2/4b3/2P1B3/4KPP1/8
Mate in 4 moves by the pawn on g2
Author's solution:
1. Nxc6+ Bxc6 2. f4+ Ke4 3. Nf6+ Kxf5 4. g4#
All program's I've tested stop once they found f4#, or another mate-in-one. But I'd like to 'force' the
solver to go on for the full 4 moves.
I think -- and this is what I'd like to test more extensively -- that if I can force a stated-depth solution
I can post-process that full solution for the conditional solution(s).
In this example, it would be something like this
a) Any variation in less than #4 gets removed
b) any 4#-solution that doesn't end in a pawn mate gets removed.
c) any 4# pawn-mate solution that is not by the g2 pawn gets removed.
Whatever remains should be a solution. (If I want to find short solutions as well, I just modify the filtering.)
But for that to work, I need to find a solver that can solve in stated depth at least. Modifying one is a possibility, but is likely to be more problematical ... | | (2) Posted by François Labelle [Tuesday, Oct 15, 2019 15:43] | Maybe this Popeye input does what you want?
begin
forsyth 1r2S3/4SQ2/2p5/qp2kP2/4b3/2P1B3/4KPP1/8
stip exact-#4
option Variation
end
The output contains the author's solution in one of the branches:
1.Se7*c6 + !
1...Be4*c6
2.f2-f4 +
2...Ke5-e4
3.Se8-f6 +
3...Ke4*f5
4.g2-g4 # | | (3) Posted by [Tuesday, Oct 15, 2019 16:16]; edited by [19-10-15] | Thanks! That looks just like what I'm looking for!
P.S. Anyone thinking that Popeye could do with a 'Dummies guide to Popeye'? :-) | | (4) Posted by James Malcom [Tuesday, Oct 15, 2019 16:45] | I most certainly think that there needs to be one! I know squat about code, and when I tried Popeye, all I ever saw was a sea of mysterious folders. That’s just me though. | | (5) Posted by seetharaman kalyan [Wednesday, Oct 16, 2019 12:55] | For dummies like me, it is Olive. | | (6) Posted by Hannu Harkola [Wednesday, Oct 16, 2019 13:45]; edited by Hannu Harkola [19-10-16] | I prefer Teddy. | | (7) Posted by James Malcom [Wednesday, Oct 16, 2019 14:24] | I could use a Dummy's Guide to any chess problem solving program! | | No more posts |
MatPlus.Net Forum Internet and Computing Problem solvers suitable for conditional problems? |
|
|
|